Abstract: Previous studies of returns to education have paid little attention to the distinction between effects of secondary vocational education and regular senior secondary school education. Using human capital and information economics theories, this study puts forward three hypotheses on the differential returns between secondary vocational school and regular senior secondary school graduates. They are tested with the data from the China Family Panel Studies in 2010 (CFPS2010). Ordinary Least Squares (OLS) often encounters problems of omitted variable and sample selection bias. Some factors affecting the choice of education are the same factors affecting returns. If these effects are not controlled, a biased coefficient could occur. Treatment effect model, with its ability to better control the effects of education choices, is better suited to estimate the differential returns of these two different types of education. The empirical results of OLS and treatment effect regression in this study show that vocational education has higher returns compared to regular high school education when education selection bias is controlled. Moreover, returns to vocational education vary across school cohorts. They appeared to have diminished since the 1990s after the initial high-return period in the early stage of the economic reform. The limited supply of college graduates in the 1980s and the 1990s higher education expansion have both contributed to this phenomenon.

Key words: CFPS, secondary vocational education, treatment effectmodel, regular senior secondary school, returns to education